storage controller: use proper ScheduleContext when evacuating a node #9908
+231
−121
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jcsp
force-pushed
the
jcsp/storcon-context-iterator
branch
from
66f7e23
to
d70d741
Compare
VladLazar
approved these changes
Nov 27, 2024
|
|
||
| if tenant_shard_id.shard_number.0 == tenant_shard_id.shard_count.count() - 1 { | ||
| let tenant_id = tenant_shard_id.tenant_id; | ||
| let tenant_shards = std::mem::take(&mut tenant_shards); |
There was a problem hiding this comment.
nit: you could break tenant_shards instead which is a bit more idiomatic imo
| /// tenant while considering the individual shards within it. This iterator is a helper | ||
| /// that gathers all the shards in a tenant and then yields them together with a ScheduleContext | ||
| /// for the tenant. | ||
| struct TenantShardContextIterator<'a> { |
There was a problem hiding this comment.
nit: I'd move this into a separate module. service is already plenty big.
| @@ -305,7 +305,7 @@ impl std::ops::Add for AffinityScore { | |||
|
|
|||
| /// Hint for whether this is a sincere attempt to schedule, or a speculative | |||
| /// check for where we _would_ schedule (done during optimization) | |||
| #[derive(Debug)] | |||
| #[derive(Debug, Clone)] | |||
Comment on lines
7097
to
7100
| /// When making scheduling decisions, it is useful to have the ScheduleContext for a whole | ||
| /// tenant while considering the individual shards within it. This iterator is a helper | ||
| /// that gathers all the shards in a tenant and then yields them together with a ScheduleContext | ||
| /// for the tenant. |
There was a problem hiding this comment.
Did you mean to delete this?
6941 tests run: 6633 passed, 0 failed, 308 skipped (full report)Code coverage* (full report)
* collected from Rust tests only The comment gets automatically updated with the latest test results
d70d741 at 2024-11-27T15:18:39.899Z :recycle: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Problem
When picking locations for a shard, we should use a ScheduleContext that includes all the other shards in the tenant, so that we apply proper anti-affinity between shards. If we don't do this, then it can lead to unstable scheduling, where we place a shard somewhere that the optimizer will then immediately move it away from.
We didn't always do this, because it was a bit awkward to accumulate the context for a tenant rather than just walking tenants.
This was a TODO in
handle_node_availability_transition:This is a precursor to #8264, where the current imperfect scheduling during node evacuation hampers testing.
Summary of changes
handle_node_availability_transitionto apply proper anti-affinity during node evacuation.